Materials Science and Engineering C 40 (2014) 324–335

Contents lists available at ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Mechanically-reinforced electrospun composite silk fibroin nanofibers containing hydroxyapatite nanoparticles Hyunryung Kim a, Lihua Che b, Yoon Ha b, WonHyoung Ryu a,⁎ a b

School of Mechanical Engineering, Yonsei University, Seoul 120-749, Republic of Korea Department of Neurosurgery, College of Medicine, Yonsei University, Seoul 120-749, Republic of Korea

a r t i c l e

i n f o

Article history: Received 2 November 2013 Received in revised form 29 March 2014 Accepted 3 April 2014 Available online 12 April 2014 Keywords: Silk fibroin Hydroxyapatite Electrospinning Composite scaffold Mechanical strength

a b s t r a c t Electrospun silk fibroin (SF) scaffolds provide large surface area, high porosity, and interconnection for cell adhesion and proliferation and they may replace collagen for many tissue engineering applications. Despite such advantages, electrospun SF scaffolds are still limited as bone tissue replacement due to their low mechanical strengths. While enhancement of mechanical strengths by incorporating inorganic ceramics into polymers has been demonstrated, electrospinning of a mixture of SF and inorganic ceramics such as hydroxyapatite is challenging and less studied due to the aggregation of ceramic particles within SF. In this study, we aimed to enhance the mechanical properties of electrospun SF scaffolds by uniformly dispersing hydroxyapatite (HAp) nanoparticles within SF nanofibers. HAp nanoaprticles were modified by γ-glycidoxypropyltrimethoxysilane (GPTMS) for uniform dispersion and enhanced interfacial bonding between HAp and SF fibers. Optimal conditions for electrospinning of SF and GPTMS-modified HAp nanoparticles were identified to achieve beadless nanofibers without any aggregation of HAp nanoparticles. The MTT and SEM analysis of the osteoblasts-cultured scaffolds confirmed the biocompatibility of the composite scaffolds. The mechanical properties of the composite scaffolds were analyzed by tensile tests for the scaffolds with varying contents of HAp within SF fibers. The mechanical testing showed the peak strengths at the HAp content of 20 wt.%. The increase of HAp content up to 20 wt.% increased the mechanical properties of the composite scaffolds, while further increase above 20 wt.% disrupted the polymer chain networks within SF nanofibers and weakened the mechanical strengths. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Silk fibroin (SF) has been investigated as a promising biomaterial in various biomedical applications including surgical sutures [1], membranes [2], drug delivery [3], and tissue engineering [4]. The outstanding biocompatibility and potential to replace collagen of extracellular matrix (ECM) offer much advantage over other synthetic biomaterials [5–7]. Recently, SF has also demonstrated promising performance as scaffold material for regenerative tissue engineering and it has been employed to construct scaffolds for various tissues such as bone [8,9], cartilage [10], ligament [11], tendon [12], and blood vessels [13]. In particular, for bone tissue engineering, SF gained much interest as a scaffold material and various strategies were developed to create a three-dimensional (3D) SF structure with high porosity and osteoconductivity. Salt-leaching or freeze-drying methods were used to create 3D SF porous scaffolds [14–16]. Furthermore, in attempts to mimic bone ECM comprised of organic components such as collagen as well as inorganic component such as hydroxyapatite, the incorporation of

⁎ Corresponding author. Tel.: +82 2 2123 5821. E-mail address: [email protected] (W. Ryu).

http://dx.doi.org/10.1016/j.msec.2014.04.012 0928-4931/© 2014 Elsevier B.V. All rights reserved.

the ceramic component into SF scaffolds has also been demonstrated [8, 15,16]. Fibrous SF scaffolds were also constructed using electrospinning technique and the bone regeneration of the electrospun SF scaffolds was demonstrated [8,17,18]. Although electrospun scaffolds were favored in many investigations because of their close resemblance to natural tissues such as collagen and high interconnectivity, the fibrous structure is significantly weak to physical loadings compared to porous scaffolds. Thus, in order to use fibrous SF scaffolds to sustain tensional or compressional stresses for skeletal tissue engineering, enhancing the mechanical properties of scaffolds is much desired [19–21]. For the increase of mechanical strength, most composite SF fibrous scaffolds have been fabricated by depositing or growing ceramics on the surface of electrospun SF nanofibers [22–25]. Although incorporation of osteo-conductive mineral components such as HAp nanoparticles within single polymer fibers was previously reported for PLA or PCL/gelatin nanofibers [26,27], there has been relatively a limited number of investigations for SF nanofibers containing HAp nanoparticles. In a previous study, inclusion of HAp nanoparticles in SF nanofibers was reported [8] but more systematic study is required to understand the effect of HAp concentrations on the HAp distribution and the degree of enhancement in mechanical strength.

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

The most significant issue of incorporation HAp nanoparticles in the SF matrix is HAp agglomeration and non-uniform distribution due the hydrophilic nature of the HAp nanoparticles and hydrophobic SF polymer. Recent studies confirmed the aggregation issue of HAp particles in SF/HAp composite scaffolds [8,9,15,22]. Such aggregation of HAp particles in the SF phase caused the weakening of mechanical strengths [28,29] and the resulting non-uniform physical and chemical properties of scaffolds due to the aggregation are less desirable. Various attempts to enhance dispersion of HAp particles in polymer matrices have been investigated. HAp nanoparticles were incorporated in poly(ε-caprolactone) by using oleic acid, an amphiphilic surfactant [30]. In other investigations, HAp surfaces were modified using stearic acid [31], carboxylic acids [32], or sodium dodecyl sulphate [33]. Dodecyl alcohol was also used to esterify HAp particles, while this was for improving the dispersion only in ethanol as well as changed some chemical properties of HAp [34]. However, these surfactants are either immiscible in aqueous SF solution or have never been applied to SFbased scaffolds yet. In another study, HAp particle dispersion was enhanced by buffering the HAp and SF solution at pH 6.8. However, such approach was demonstrated only for low HAp concentrations [8] and was difficult to avoid HAp aggregation for higher HAp concentrations. Thus, in this study, we aimed to investigate how much improvement in the mechanical properties of SF electrospun scaffolds could be achieved by adjusting the distribution and content of HAp within silk nanofibers. For uniform distribution of HAp particles in silk matrix, a silane coupling agent, GPTMS, was incorporated in the preparation of

325

SF/HAp solution for electrospinning [35–37]. The electrospinning parameters were also optimized for the composite silk solution containing GPTMS-modified HAp nanoparticles. The surface morphology and the HAp dispersion of the electrospun composite fibers were analyzed using SEM and TEM images. The mechanical properties of composite silk electrospun scaffolds were investigated using tensile tests. Finally, the cell toxicity and the proliferation of the composite scaffolds were assessed by culturing human osteoblasts on the silk scaffolds with varying content of HAp. 2. Experimental 2.1. Materials Bombyx mori silkworm cocoons were purchased from the Uljin farm, Rep. of Korea. Sodium Oleate (37225-01) was purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). Sodium carbonate Anhydrous (7541-4405) was supplied by Daejung Chemicals & Metals Co., LTD, Rep. of Korea. Lithium bromide anhydrous (L1106) was provided from Samchun Chemical Inc., Rep. of Korea. Nano HAp (nHAp) powder was generously donated by BioAlpha Inc., Rep. of Korea. GPTMS was supplied from TCI (Tokyo, Japan). Poly(ethylene glycol) (PEG) (Mw = 10,000) and poly(ethylene oxide) (PEO) (Mw = 900,000) were purchased from Sigma (St. Louis, MO, USA). Human osteoblast cell line (hFOB1.19, CRL-11371™) was purchased from ATCC and collagen sponge (AteloPlug™, Bioland, OChang, Rep. of Korea) was donated by

(a)

(b)

(c)

(d)

Fig. 1. SEM images of electrospun SF/HAp scaffolds (a) using parameters as in the previous study by Li et al. [8] (SF:PEO = 82:18 (w:w), 20% HAp suspension of 0.1125 ml, and electric field strength of 12 kV/21.5 cm, (b) with 2.75% HAp suspension of 0.8 ml under 12 kV/18 cm, and (c) with SF:PEO = 75:25 (w:w) and 10% HAp suspension of 0.5 ml under 14 kV/15 cm. (d) TEM image of HAp nanoparticles.

326

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

Fig. 2. (a) Schematic diagram of anhydrous surface modification of HAp nanoparticles by GPTMS. (b) Enhancement of HAp dispersion within SF nanofibers after GPTMS treatment on HAp nanoparticles.

BioAlpha Inc., Rep. of Korea. DMEM/F-12 without phenol red and penicillin–streptomycin were provided from Gibco (Carlsbad, CA, USA). Hyclone fetal bovine serum was supplied by Thermo Scientific (Waltham, MA, USA). G418, Pyrex ® cloning cylinder (O.D. × H10 mm × 10 mm), and 3-(4,5-dimthyldiazol-2-yl)-2,5,-diphenyl tetrazolium bromide (MTT) were purchased from Sigma (St. Louis, MO, USA). 12-well plates were provided from Nunc (Penfield, NY, USA). DMSO solution and OsO4 were supplied by Georgia Chem (Norcross, GA, USA) and Polysciences, Inc. (Warrington, PA, USA), respectively.

40 wt.% was mixed with 2.5 ml of aqueous SF solution (9 wt.%). An aqueous solution of poly(ethylene oxide) (Mw = 900,000) at 5 wt.% was prepared by dissolving PEO in distilled water and stirring for 5 days at room temperature. Afterwards, 1.5 ml of the PEO solution was added to the 2.5 ml the GPTMS-HAp added SF solution (SF:PEO = 3:1 (w/w)) to adjust viscosity for spinning stability.

2.2. Preparation of SF aqueous solution Cocoons of B. mori were degummed with 0.2 M of Na2CO3 and 0.01 M of sodium oleate aqueous solution, boiled twice and rinsed thoroughly with distilled water to remove sericin proteins from raw silk fiber. The degummed SF fibers were dried for 24 h at room temperature and dissolved in 9.3 M LiBr solution at 60 °C, resulting in 15% (w/v) solution. This solution was dialyzed against distilled water using a dialysis tubing membrane (Mw 11,252), which yielded approximately 5 wt.% SF aqueous solution. This solution was concentrated by dialyzing against 25 wt.% PEG (Mw 10,000) solution, resulting in 9 wt.% SF solution. 2.3. Surface modification of HAp using GPTMS HAp nanoparticles were surface-modified using GPTMS by following a method described by Li et al. [35] to improve HAp dispersion in SF solution. Briefly, HAp nanoparticles were dried at 120 °C for 24 h to remove moisture before the GPTMS treatment. HAp was mixed in 2% (v/v) GPTMS–ethanol solution, which was stirred at 120 rpm under 60 °C for 2 h. The solution was further heated at 80 °C for 2 h to evaporate ethanol in convection oven. GPTMS-HAp was dried in vacuum oven under 120 °C for 24 h. Chemical composition of HAp and GPTMS-HAp pressed into the KBr pellets was analyzed using FT-IR spectrometer (Vertex70, Bruker) in the range of wavenumbers from 500 to 1500 cm−1 in transmittance mode at room temperature. 2.4. Preparation of electrospinning solution GPTMS-modified HAp was suspended in distilled water to yield aqueous suspensions with the concentrations from 10, 20, 30, to 40 wt.%. Each suspension was sonicated for 5 min. Then, 0.225 ml of aqueous suspension of GPTMS-HAp with the concentration of 0 to

Fig. 3. (a) FTIR of HAp and GPTMS-modified HAp. Peaks at 1200 and 1095 cm−1 indicate Si\OCH3 bonds and the peak at 908 cm−1 indicates epoxide bond. (b) Sedimentation test of HAp/SF and GPTMS-modified HAp/SF solution (12 h after mixing).

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

(a)

(b)

(c)

(d)

327

Fig. 4. (a) SEM and (b) TEM images of SF/HAp composite scaffolds (HAp/SF = 10 wt.%) without surface treatment. (c) SEM and (d) TEM images of SF/HAp composite scaffolds (HAp/SF = 10 wt.%) after surface treatment by a coupling agent GPTMS.

2.5. Viscosity measurement of HAp/SF mixture Each mixture was transported into the cup which was subsequently installed to the viscometer (DV2TLV, Brookfield). Spindle (CP41) was rotated at 1 rpm for 20 s and yielded 10 data points, each of which was measured in every 2 s, and averaged. All viscosities were measured under room temperature. 2.6. Electrospinning A commercial electrospinning system (Model ESR200RD, Nano NC Inc., Seoul, Rep. of Korea) was used in this study. Each syringe was filled with 4 ml of spinning solution and the solution was pumped at a feed rate of 1.2 ml/h using a syringe pump. The metal needle (17 gauge) of the syringe connected to a high voltage power supply. A drum collector covered with aluminum foil was rotated at 500 rpm and grounded such that the ejected solution from the needle tip could be collected on it. The electric field strength was kept at 14 kV/15 cm. The electrospun SF scaffold was methanol-treated for 5 min and water vapor-treated for 12 h before the cell culture to induce β-sheet formation. 2.7. SEM, EDS, and TEM analysis Electrospun SF scaffold was sputter-coated with Pt and the morphology and the element distribution were analyzed with JEOL-7001 F Field Emission SEM (JEOL, Tokyo, Japan). JEM-2010 TEM (High Contrast) (JEOL, Tokyo, Japan) was operated 200 kV to observe HAp particles and at 120 kV to examine the distribution of nHAp particles embedded on the SF scaffold which was placed directly on the carbon grid. For the measurement of fiber diameters, three different spots from each SEM

image were randomly selected and 100 fibers were randomly chosen for each condition of HAp content. 2.8. Tensile tests of hybrid scaffolds Tensile test of SF/HAp hybrid scaffolds was conducted based on ASTM D882 standard test method using Universal Testing Machine (WL2100, Withlab Co., LTd., Rep of Korea). The dimension of the specimen was 12 mm in width and 200 mm in height. The testing was done at the speed of 50 mm/min at room temperature with a load cell capacity of 200 N. Tensile strength, yield strength, and Young's modulus were measured in this test. 2.9. Statistical analysis The data collected were expressed as mean ± standard deviation (S.D.). Significant outliers from the data were excluded using Grubb's test. All the data were analyzed with Student's t test (p b 0.05). 2.10. Cell culture The cell culture medium consisted of 89% DMEM/F-12 without phenol red, 10% fetal bovine serum, 1% penicillin–streptomycin, and Table 1 Viscosity of HAp–SF mixture with varying HAp content under room temperature.

Viscosity (cP) Standard deviation

0%

10%

20%

30%

40%

205.57 ±1.19

276.36 ±4.73

285.68 ±4.65

350.71 ±7.00

470.42 ±2.02

328

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

0.3 mg/ml G418. Human osteoblast cells were cultured in an incubator at 37 °C under 5% CO2 concentration. Sub-culturing of the cells was performed at a ratio of 1:4 every 4 or 5 days. The prepared scaffolds were sterilized in 100% ethanol over 12 h, washed three times with PBS for 30 min each time at room temperature, and neutralized in complete medium overnight at 37 °C in 5% CO2 incubator. In order to hold scaffolds at the bottom of each well plate, the autoclaved Pyrex ® cloning cylinder (O.D. × H10 mm × 10 mm) was placed on the surface of the scaffolds. 5 × 104 numbers of cells in 10 μl of complete medium were inoculated on the surface of the pretreated scaffolds inside the cloning cylinders. The well plates containing these cylinders were kept at a room temperature for 20 min and 2 ml of complete medium was added into each well. After incubation of the well plate in the CO2 incubator, the medium was changed every three or four days by removing 1 ml of cell culture media and carefully adding 1 ml of fresh complete medium.

2.12. SEM analysis of cultured cells

2.11. MTT assay

3.1. Electrospinning of SF/HAp mixture

By the end of day 14, samples (n = 3 per group) in the cell culture media were supplemented with 0.5 mg MTT/ml, and incubated in the dark at 37 °C, 5% CO2 for 4 h. The scaffolds were transferred to a clean 14 ml polystyrene tubes containing 500 μl DMSO solvent. Scaffolds with cells were ground using pellet pestle® motor (Kontes) and 1.5 ml Pestle (Kimble chase, 749521-1500) until the scaffold color became similar to the solution color. The concentration of the supernatant was measured via UV absorption at 550 nm. Statistically significant differences of the subgroups were determined using Student's t-test. (**p b 0.01).

Several sets of parameters were examined to find ideal conditions for stable electrospinning. When the electrospinning was conducted using the conditions from the previous study by Li et al. [8], 5 wt.% (HAp/SF) scaffold showed irregular fiber thickness (Fig. 1a). Under this electrospinning condition, aggregation of HAp particles at micro scale was also often found in the electrospun scaffolds. Thus, in order to enhance HAp dispersion, the aqueous spinning solution was diluted by increasing the water amount in the HAp suspension. However, this dilution caused reduction in the viscosity of the solution, resulting in

On the 14th day after cell seeding, samples (n = 3 per group) were washed once with PBS and fixed in Karnovsky fixing solution (2% glutaraldehyde, 2% paraformaldehyde, and 0.5% CaCl2 in 0.1 M PBS buffer at pH 7.4) over 6 h. The scaffolds were washed in 0.1 M PBS for 2 h and then fixed in 1% OsO4 in 0.1 M PBS buffer for 1.5 h. OsO4 was removed by washing in 0.1 M PBS buffer for 10 min. The scaffolds were dehydrated in the gradational alcohol solutions (50%, 60%, 70%, 80%, 90%, 95%, 100%), infiltrated with the Iso-amyl acetate, and subjected to Critical Point Dryer (HCP-2, Hitachi, Japan). Lastly, it was coated with a 30 nm-thick gold layer by using Ion coater (Eiko, IB-3, Japan) for SEM analysis. 3. Results

(a) 10%

(b) 20%

(c) 30%

(d) 40%

Fig. 5. SEM images of SF/HAp composite scaffolds with increasing concentrations of GPTMS-HAp of (a) 10, (b) 20, (c) 30, and (d) 40 wt.%. (e) SF fiber distribution for each HAp concentration.

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

329

Fig. 5. (continued).

bead formation of electrospun fibers (Fig. 1b). In order to find out optimal parameters for stable electrospinning of beadless and uniform nano fibers without HAp aggregation, several parameters such as the viscosity of the solution, electric field strength, feed rate, and relative humidity were adjusted and a silane coupling agent, GPTMS, was included. As shown in Fig. 1c, the composite electrospun scaffolds had uniform fiber morphology without bead and HAp agglomeration. 3.2. GPTMS treatment of HAp nanoparticle for uniform dispersion in SF fibers A silane coupling agent, GPTMS was used to achieve more uniform dispersion of the HAp nanoparticles within each SF nano fiber. The GPTMS was covalently bonded onto the surface of HAp and made HAp surface more dissolvable in SF polymer chain networks (Fig. 2). In the FTIR spectrum of GPTMS-HAp (Fig. 3a), the presence of additional

characteristic epoxide bond at 908 cm−1 and Si\OCH3 bonds at 1200 cm−1 and 1095 cm−1 compared to the native HAp spectrum confirms the presence of GPTMS on the HAp surface. Sedimentation test was also conducted for the GPTMS-HAp/SF mixture compared to the HAp/SF mixture (Fig. 3b). Each group contained various amounts of HAp from 10 wt.% to 40 wt.%. Both solutions showed uniform dispersion immediately after the mixing. After 12 h, both mixtures maintained overall uniform dispersion of HAp particles with minor sedimentation at the bottom of the containers. As shown in Fig. 4 a&b, when raw Table 2 Fiber diameters of fibrous scaffold with different ratios of SF and HAp.

Diameter (nm) Standard deviation

0%

10%

20%

30%

40%

463.01 ±8.71

491.73 ±7.69

480.81 ±7.29

496.76 ±9.90

469.03 ±8.49

330

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

HAp nanoparticles were used, the SF/HAp composite fibers showed the rough morphology of nanofibers due to highly-aggregated HAp nanoparticles within SF fibers. When the GPTMS-modified HAp nanoparticles were used, HAp nanoparticles were uniformly dispersed in SF nanofibers without any aggregation (Fig. 4c&d).

Humidity was kept around 40% since fibers were thinned below 40% and bead formation occurred above 40% humidity. On the other hand, further increase of HAp above 40% in the composite solution caused the particle sedimentation which subsequently resulted in the clogging problem at the needle orifice.

3.3. Optimization of electrospinning parameters

3.4. Morphology of electrospun composite nanofibers

Conditions for stable jetting were identified by varying some parameters during electrospinning. The concentrations of HAp suspension and PEO solution were adjusted in a way that enabled stable electrospinning even at high HAp content up to 40% compared to the SF amount. First, PEO content in a SF electrospinning solution (18/82 → 25/75, PEO/SF, w/w) was increased to minimize bead formation. It was found that viscosity less than 205 cP yielded unstable jetting (Table 1). Then, in order to avoid the HAp aggregation, a silane coupling agent, GPTMS, was included to enhance the dispersion of relatively hydrophilic HAp nanoparticles in hydrophobic SF polymer chain networks. When HAp content increased up to 40% in an electrospinning solution, at an electric field of 14 kV/15 cm, Taylor cone became less stable. The field strength had to be reduced down to 12 kV/15 cm to achieve stable electrospinning, although the viscosity of the solution increased from 206 cP to 470 cP (Table 1). The feed rate was set at 1.2 ml/h such that the feed rate balanced the evaporation rate. Electrospinning solution was clogged at feed rates lower than 1.2 ml/h due to faster evaporation than the solution feeding at the nozzle, while it dripped at feed rates higher than 1.2 ml/h due to too much solution feeding than electrical drawing by an electrical field.

The electrospun SF–HAp nanofiber composites with increasing HAp contents from 0 to 40% (HAp/Silk, w/w) were examined by using SEM (Fig. 5). SF fibers were randomly distributed throughout the scaffold with uniform fiber diameters. The fiber diameters of the composite scaffolds were measured on 100 different fibers of each scaffold by using NIH Image J program and averaged, as shown in Table 2. It is notable that the fiber diameters of the scaffolds with different HAp concentrations showed a similar diameter of about 480 nm. This indicates that the inclusion of the HAp has no effect on the fiber diameter. The surface of SF scaffolds without HAp or low HAp concentrations was mostly smooth (Fig. 5a&b). When the HAp content increased about 20%, slight aggregation of HAp particles started to appear (Fig. 5c&d). On the other hand, as shown in Fig. 2a, composite scaffolds without GPTMS treatment on HAp showed much rougher morphology even at low HAp concentration due to more severe aggregation of HAp particles within SF fibers. TEM analysis also showed that at HAp concentrations of 10 and 20%, no particle aggregation was observed (Fig. 6a&b). The HAp particles were mostly embedded inside the fibers while at only a few locations they were attached or pinned on the surface. However, as HAp concentration increased above 30 wt.% (HAp/Silk), the HAp particles

(a) 10%

(b) 20%

(c) 30%

(d) 40%

Fig. 6. TEM images of SF/HAp composite scaffolds with different mixing ratios of GPTMS-HAp to SF, (a) 10, (b) 20, (c) 30, (d) 40 wt.%.

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

331

Fig. 7. EDS mapping of Ca element in SF/HAp composite scaffolds with increasing concentrations of HAp from (a) 10, (b) 20, (c) 30, and (d) 40 wt.%.

started forming agglomeration (Fig. 6c&d). In addition, as shown in Fig. 7, EDS mapping of Ca element (a major component of HAp) showed increasing amount of Ca element in the composite scaffolds electrospun from the SF/HAp solution with higher HAp concnetrations.

osteoblasts stretched fully and grew on the smooth surfaces of collagen scaffolds, indicating high proliferation and biocompatibility. Osteoblast culture on the electrospun SF/HAp scaffolds also showed similar morphology (Fig. 9b).

3.5. Cell culture on composite scaffolds

3.6. Mechanical properties of hybrid scaffolds

Cytotoxicity of the scaffolds with different HAp concentrations was evaluated using MTT assay after 14 days of cultivation of osteoblasts. Fig. 8 shows the absorbance of MTT crystal dissolved in DMSO solution. The MTT assay result indicates that SF/HAp scaffolds showed about 45% of bioactivity compared to a collagen control group. Although composite scaffolds with 10% of HAp content showed slightly higher proliferation of osteoblasts, there was no significant correlation between cell proliferation and HAp content in the composite scaffolds. As shown in Fig. 9a,

Tensile tests were performed with the composite scaffolds of silk fibroin and nHAp particles. Electrospun samples with different HAp contents of between 0 and 40 wt.% (HAp/Silk) were prepared in a form of 12 mm width and 200 mm length. During the tensile tests, tensile strengths, yield strengths, modulus, and maximum strain were measured for each sample. Interestingly, all those mechanical properties (except maximum strain) increased until HAp concentration became 20% and then decreased at higher HAp contents (Fig. 10a–c). Unlike other properties, the maximum strain of the composite scaffolds became smaller as they contained a larger amount of HAp (Fig. 10d). 4. Discussion

Fig. 8. MTT assay for osteoblast activity on collagen sponge and electrospun SF/HAp scaffolds with varying HAp contents. The absorbance was measured at 550 nm.

Electrospinning of a combination of liquid and particulate-based suspension often results in poorly-controlled nanofiber structures. In previous works, electrospinning of liquid–liquid or liquid–inorganic particles has been investigated to achieve uniform nanofibers [38,39]. In this study, we developed electrospinning of a mixture of liquid SF solution and HAp nanoparticles using a coupling agent which prevented aggregation of ceramic particles. This allowed for electrospinning of uniform nanofibers without aggregated HAp even with a single injection nozzle. It was shown that the fiber diameters of the scaffolds were similar regardless of the different HAp contents (Table 2). However, in general, higher concentration increases viscosity and this leads to the less stretching of the polymer solution and an increase in the diameter of electrospun fibers. It seems that the inclusion of HAp had less effect on the stretching of the polymer solution during the flight time of electrospinning. Although the stretching of polymer solution is mostly governed by the solution viscosity, the “undeformable solid” HAp does

332

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

not participate in the stretching of polymer solution and this may explain the minimal effect of the HAp content on the fiber diameter. Uniform dispersion and enhancing interfacial bonding between HAp nanoparticles and SF are desired to improve the mechanical properties of electrospun nanofibers. A previous study demonstrated uniform dispersion of HAp nanoparticles in electrospun nanofibers by buffering SF solution and HAp suspension at pH 6.8 [8]. However, this approach has not improved the HAp distribution when higher concentration of HAp was used in our experiments. As mentioned in the previous study, appropriate combination of sonication and magnetic stirring was required and it became more difficult with higher concentration of HAp nanoparticles. In addition, during electrospinning over 30 min, HAp sedimentation often occurred. On the other hand, hydroxyl groups on HAp surfaces can be modified by GPTMS to have relatively

hydrophobic epoxide rings exposed to the outside [36]. It was also reported that the epoxide end of GPTMS reacted with the side chains of silk fibroin, such as lysine, histidine, arginine, and tyrosine, and contributed to the mechanical strengths of SF/HAp fibers [37]. Thus, GPTMS-treated HAp nanoparticles were expected to enhance an interfacial bonding between HAp nanoparticles and SF polymer chains. Although the use of GPTMS improved the HAp particle distribution within the fibers, the phase separation and aggregation of the HAp particles still occurred in the electrospinning solution. The precipitation often hindered stable electrospinning and resulted in uncontrolled and less uniform fibrous structures. This was probably due to the presence of a water layer covering the surface of HAp particles due to incomplete HAp drying before the reaction between GPTMS and HAp particles. The water molecules on the HAp surface, which were not

(a)

(b)

Fig. 9. SEM images of osteoblasts cultured on (a) collagen and (b) SF/HAp composite scaffold (HAp/SF = 10 wt.%).

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

completely removed from the drying, prevented or weakened the interaction between the coupling agent and HAp [40]. In general, GPTMS functioned properly as a chemical intermediate that promoted interaction between aqueous silk solution and HAp particles. However, it is required to avoid any exposure to humidity or incomplete drying of the HAp particles before surface treatment. It was demonstrated that SF–HAp composite scaffolds showed the maximum mechanical properties at 20 wt.% of HAp. Compared to the scaffolds without HAp, scaffolds with HAp showed higher mechanical strength since the HAp particles functioned as “filler particles” and enhanced matrix yielding [41]. This resulted in transition from flexible to tough behavior. The initial enhancement of mechanical properties is attributed to the stiffness of HAp particles that reduced the flexibility

333

of SF. However, when the amount of HAp particles exceeded a critical volume fraction for aggregation, HAp particles introduced disruption and functioned as discontinuities among fiber chains [41] because the distance between the particles decreased and the toughness substantially decreased (Fig. 10e). This is because the holes (or defects) around HAp particles derived from the stress concentration under tensile readily amalgamate with adjacent holes [42]. Thus, an increase of HAp content in silk fibers directly caused such accelerated breakage of scaffolds during tensile tests. The critical volume fraction in this study seemed to be formed around at 20 wt.% of HAp and the scaffolds exhibited a brittle mode with HAp content above 20 wt.%. This also could explicate the reason the maximum strain dropped as more HAp was added to the scaffold. HAp particles acted as a source of holes and the

Fig. 10. Effects of HAp contents on (a) tensile strength, (b) yield strength, (c) Young's modulus, and (d) maximum strain of SF/HAp composite scaffolds. (e) Models of HAp particle behavior at low (mode I) and high (mode II) concentrations under tensile stress.

334

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335

presence of higher HAp content induced narrowing of the gap between holes. Thus, with higher HAp contents, neighboring holes were more likely to coalesce into macro fracture quickly. Biocompatibility of SF scaffold was evaluated based on the viability of osteoblasts using MTT assay analysis and SEM imaging of the morphology of the cells cultured on the surface of the hybrid scaffolds. No distinctive biological advantage of incorporating HAp in SF was exhibited. It is ascribed to the fact that HAp particles were embedded within SF fibers and had no direct contact with the cells to influence the cell proliferation. Architectural modification for SF scaffold such as pore size and fiber diameter also seems necessary for osteoblasts to favorably attach and grow on the hybrid scaffolds to substitute collagen sponge. SF scaffolds from aqueous SF solution were reported to degrade in vivo after 2–6 months after their implantation [43]. SF scaffolds dissolve slowly in water depending on the degree of crystallinity (e.g. beta sheet stacking). After dissolution, enzymatic degradation breaks down each protein chain. Since it is comprised of amino acids, there is usually no hazardous material produced from the degradation process. However, since the absorption rate is dependent on the morphological and structural features of the SF scaffolds as well as the environment of their implant sites, it is also required to perform the in vivo absorption study to assess the feasibility of SF scaffolds for clinical applications. Further study on the biocompatibility of HAp, GPTMS, and scaffold porosity as well as bioabsorption is essential to improve this research. 5. Conclusions SF solutions containing HAp nanoparticles at concentrations up to 40 wt.% was electrospun to fabricate SF/HAp composite scaffolds for bone regeneration. The SF/HAp ratios were varied to investigate the effect of HAp content on the morphology, mechanical properties, and biocompatibility of the composite scaffolds. The surface of the SF/HAp nano fibers became rougher with the increase of HAp content. Use of a coupling agent, GPTMS, enabled more uniform distribution of HAp particles within the SF nano fibers. Tensile testing of the composite scaffolds showed the peak strength at 20% of HAp. HAp nanoparticles below 20% reinforced SF polymer chains to bear high load. However, at higher HAp content than 20%, the HAp particles disrupted the SF polymer chains, which resulted in an adverse effect on the mechanical properties. Cell cultivation and MTT assay demonstrated that SF/HAp composite scaffolds were biocompatible to osteoblasts. Acknowledgments This research was supported by the Small & Medium Business Administration funded by the Ministry of Knowledge Economy (000451390112). Yonsei University Institute of HRD Program for Nano/Micro Mechanical Engineering, a Brain Korea 21 program, Korea also provided the financial aid. Slurry of HAp nanoparticles was generously donated by BioAlpha Inc. (Sungnam, Rep. of Korea). Mechanical tensile test was done by the Korea Polymer Testing & Research Institute (KOPTRI), Ltd. The authors thank Sung Yeun Yang and Tae Heon Hwang for instruction about silk fibroin and electrospinning. References [1] G.H. Altman, F. Diaz, C. Jakuba, T. Calabro, R.L. Horan, J. Chen, H. Lu, J. Richmond, D.L. Kaplan, Silk-based biomaterials, Biomaterials 24 (2003) 401–406. [2] H.J. Jin, J. Park, R. Valluzzi, P. Cebe, D.L. Kaplan, Biomaterial films of Bombyx mori silk fibroin with poly(ethylene oxide), Biomacromolecules 5 (2004) 711–717. [3] S. Hofmann, C.T. Wong Po Foo, F. Rossetti, M. Textor, G. Vunjak-Novakovic, D.L. Kaplan, H.P. Merkle, L. Meinel, Silk fibroin as an organic polymer for controlled drug delivery, J. Control. Release 111 (2006) 219–227. [4] U.-J. Kim, J. Park, H.J. Kim, M. Wada, D.L. Kaplan, Three-dimensional aqueousderived biomaterial scaffolds from silk fibroin, Biomaterials 26 (2005) 2775–2785.

[5] K. Inouye, M. Kurokawa, S. Nishikawa, M. Tsukada, Use of Bombyx mori silk fibroin as a substratum for cultivation of animal cells, J. Biochem. Biophys. Methods 37 (1998) 159–164. [6] Y. Yang, X. Chen, F. Ding, P. Zhang, J. Liu, X. Gu, Biocompatibility evaluation of silk fibroin with peripheral nerve tissues and cells in vitro, Biomaterials 28 (2007) 1643–1652. [7] H.-J. Jin, J. Chen, V. Karageorgiou, G.H. Altman, D.L. Kaplan, Human bone marrow stromal cell responses on electrospun silk fibroin mats, Biomaterials 25 (2004) 1039–1047. [8] C. Li, C. Vepari, H.-J. Jin, H.J. Kim, D.L. Kaplan, Electrospun silk-BMP-2 scaffolds for bone tissue engineering, Biomaterials 27 (2006) 3115–3124. [9] H.J. Kim, U.-J. Kim, H.S. Kim, C. Li, M. Wada, G.G. Leisk, D.L. Kaplan, Bone tissue engineering with premineralized silk scaffolds, Bone 42 (2008) 1226–1234. [10] Y. Wang, D.J. Blasioli, H.-J. Kim, H.S. Kim, D.L. Kaplan, Cartilage tissue engineering with silk scaffolds and human articular chondrocytes, Biomaterials 27 (2006) 4434–4442. [11] G.H. Altman, R.L. Horan, H.H. Lu, J. Moreau, I. Martin, J.C. Richmond, D.L. Kaplan, Silk matrix for tissue engineered anterior cruciate ligaments, Biomaterials 23 (2002) 4131–4141. [12] S. Sahoo, S.L. Toh, J.C.H. Goh, A bFGF-releasing silk/PLGA-based biohybrid scaffold for ligament/tendon tissue engineering using mesenchymal progenitor cells, Biomaterials 31 (2010) 2990–2998. [13] X. Zhang, C.B. Baughman, D.L. Kaplan, In vitro evaluation of electrospun silk fibroin scaffolds for vascular cell growth, Biomaterials 29 (2008) 2217–2227. [14] S. Hofmann, H. Hagenmüller, A.M. Koch, R. Müller, G. Vunjak-Novakovic, D.L. Kaplan, H.P. Merkle, L. Meinel, Control of in vitro tissue-engineered bone-like structures using human mesenchymal stem cells and porous silk scaffolds, Biomaterials 28 (2007) 1152–1162. [15] Y. Zhang, C. Wu, T. Friis, Y. Xiao, The osteogenic properties of CaP/silk composite scaffolds, Biomaterials 31 (2010) 2848–2856. [16] S. Bhumiratana, W.L. Grayson, A. Castaneda, D.N. Rockwood, E.S. Gil, D.L. Kaplan, G. Vunjak-Novakovic, Nucleation and growth of mineralized bone matrix on silk– hydroxyapatite composite scaffolds, Biomaterials 32 (2011) 2812–2820. [17] C.S. Ki, S.Y. Park, H.J. Kim, H.-M. Jung, K.M. Woo, J.W. Lee, Y.H. Park, Development of 3-D nanofibrous fibroin scaffold with high porosity by electrospinning: implications for bone regeneration, Biotechnol. Lett. 30 (2008) 405–410. [18] S.Y. Park, C.S. Ki, Y.H. Park, H.M. Jung, K.M. Woo, H.J. Kim, Electrospun silk fibroin scaffolds with macropores for bone regeneration: an in vitro and in vivo study, Tissue Eng. A 16 (2010) 1271–1279. [19] D.W. Hutmacher, Scaffolds in tissue engineering bone and cartilage, Biomaterials 21 (2000) 2529–2543. [20] G. Wei, P.X. Ma, Structure and properties of nano-hydroxyapatite/polymer composite scaffolds for bone tissue engineering, Biomaterials 25 (2004) 4749–4757. [21] Q. Fu, E. Saiz, M.N. Rahaman, A.P. Tomsia, Bioactive glass scaffolds for bone tissue engineering: state of the art and future perspectives, Mater. Sci. Eng. C 31 (2011) 1245–1256. [22] K. Wei, Y. Li, K.-O. Kim, Y. Nakagawa, B.-S. Kim, K. Abe, G.-Q. Chen, I.-S. Kim, Fabrication of nano-hydroxyapatite on electrospun silk fibroin nanofiber and their effects in osteoblastic behavior, J. Biomed. Mater. Res. 97A (2011) 272–280. [23] Y. Ito, H. Hasuda, M. Kamitakahara, C. Ohtsuki, M. Tanihara, I.K. Kang, et al., A composite of hydroxyapatite with electrospun biodegradable nanofibers as a tissue engineering material, J. Biosci. Bioeng. 100 (2005) 43–49. [24] K. Rodriguez, S. Renneckar, P. Gatenholm, Biomimetic calcium phosphate crystal mineralization on electrospun cellulose-based scaffolds, ACS Appl. Mater. Interfaces 3 (2011) 681–689. [25] C. He, G. Xiao, X. Jin, C. Sun, P.X. Ma, Electrodeposition of nanofibrous polymer scaffolds: rapid mineralization, tunable calcium phosphate composition and topography, Adv. Funct. Mater. 20 (2010) 3568–3576. [26] F. Peng, X. Yu, M. Wei, In vitro cell performance on hydroxyapatite particles/poly (L-lactic acid) nanofibrous scaffolds with an excellent particle along nanofiber orientation, Acta Biomater. 7 (2011) 2585–2592. [27] S.H. Jegal, J.H. Park, J.H. Kim, T.H. Kim, U.S. Shin, T.I. Kim, H.W. Kim, Functional composite nanofibers of poly (lactide–co-caprolactone) containing gelatin– apatite bone mimetic precipitate for bone regeneration, Acta Biomater. 7 (2011) 1609–1617. [28] H.-W. Kim, Biomedical nanocomposites of hydroxyapatite/polycaprolactone obtained by surfactant mediation J, Biomed. Mater. Res. 83A (2007) 169–177. [29] J.-H. Song, H.-E. Kim, H.-W. Kim, Electrospun fibrous web of collagen–apatite precipitated nanocomposite for bone regeneration, J. Mater. Sci. Mater. Med. 19 (2008) 2925–2932. [30] H.-W. Kim, Biomedical nanocomposites of hydroxyapatitle/polycaprolactone obtained by surfactant mediation, J. Biomed. Mater. Res. 83A (2006) 169–177. [31] Y. Li, W. Weng, Surface modification of hydroxyapatite by stearic acid: characterization and in vitro behaviors, J. Mater. Sci. Mater. Med. 19 (2008) 19–25. [32] H. Tanaka, T. Watanabe, M. Chikazawa, K. Kandon, T. Ishikawa, TPD, FTIR, and molecular adsorption studies of calcium hydroxyapatite surface modified with hexanoic and decanoic acids, J. Colloid Interface Sci. 214 (1999) 31–37. [33] S. Shimabayashi, T. Uno, M. Nakagaki, Formation of a surface complex between polymer and surfactant and its effect on the dispersion of solid particles, Colloids Surf. A Physicochem. Eng. Asp. 123–124 (1997) 283–295. [34] L. Borum-Nicholas, O.C. Wilson Jr., Surface modification of hydroxyapatite. Part I. Dodecyl alcohol, Biomaterials 24 (2003) 3671–3679. [35] L. Li, G. Li, J. Jiang, X. Liu, L. Luo, K. Nan, Electrospun fibrous scaffold of hydroxyapatite/poly (ε-caprolactone) for bone regeneration, J. Mater. Sci. Mater. Med. 23 (2012) 547–554.

H. Kim et al. / Materials Science and Engineering C 40 (2014) 324–335 [36] M.T. Arafat, C.X.F. Lam, A.K. Ekaputra, S.Y. Wong, C. He, D.W. Hutmacher, X. Li, I. Gibson, High performance additive manufactured scaffolds for bone tissue engineering application, Soft Matter 7 (2011) 8013–8022. [37] M. Tsukada, M. Nagura, H. Ishikawa, H. Shiozaki, Structural characteristics of silk fibers treated with epoxides, J. Appl. Polym. Sci. 43 (1991) 643–649. [38] G. Larsen, R. Velarde-Ortiz, K. Minchow, A. Barrero, I.G. Loscertales, A method for making inorganic and hybrid (organic/inorganic) fibers and vesicles with diameters in the submicrometer and micrometer range via sol−gel chemistry and electrically forced liquid jets, J. Am. Chem. Soc. 125 (2003) 1154–1155. [39] A. Townsend-Nichloson, S.N. Jayasinghe, Cell electrospinning: a unique biotechnique for encapsulating living organisms for generating active biological microthreads/scaffolds, Biomacromolecules 7 (2006) 3364–3369.

335

[40] Q. Liu, Hydroxyapatite/Polymer Composites for Bone Replacement, UT, Enschede, The Netherlands, 1997. [41] Y. Ou, F. Yang, Z.-Z. Yu, A new conception on the toughness of nylon6/silica nanocomposite prepared via in situ polymerization, J. Polym. Sci. B Polym. Phys. 36 (1998) 789–795. [42] M. Wang, W. Bonfield, Chemically coupled hydroxyapatite-polyethylene composites: structure and properties, Biomaterials 22 (2001) 1311–1320. [43] Y. Wang, D.D. Rudym, A. Walsh, L. Abrahamsen, H.J. Kim, H.S. Kim, C. Kirker-Head, D. L. Kaplan, In vivo degradation of three-dimensional silk fibroin scaffolds, Biomaterials 29 (2008) 3415–3428